A. Complex System

It is no longer acceptable – indeed, it is incredibly risky for Western democracies – for antitrust agencies to formulate policy without consideration of geopolitical consequences. Potential short-term domestic consumer welfare issues pale in comparison to many potentially existential geopolitical threats. Fortunately, dealing presciently with these issues will aid competition and innovation in the United States and elsewhere.

There is reason for concern. In the past, antitrust enforcement actions (including by the Federal Trade Commission (FTC)) in the global technology marketplace have been misadventures in terms of their impact on not only competition but also US global competitiveness. Many are by now aware that China is endeavoring to stack international standards bodies with individuals who swear fealty to China. “China Standard 2035” lays out objectives for blockchain, quantum computing, AI, and other technologies. Impacting and controlling 5G standards development is also part of China’s national mission. The Western democracies must be mindful of these activities and the underlying strategy. It is important to prevent the politicization of the standards development process.Footnote ⁸ The 14th Five-Year Plan for National Informatization released in December 2021, for example, restates China’s goals to “create a closed-loop innovation mechanism” to promote “standards building” in prioritized areas such as 5G, Big Data, AI, blockchain, industrial Internet, and so forth, and accelerate the completion and perfection of existing data sharing and data application standard systems. It also anticipates greater integration of China’s information standards efforts into all sectors of the Chinese economy relying upon this “closed-loop” innovation system.Footnote ⁹ Policymakers and executives and members of standards bodies must be mindful of these activities and the underlying strategy.

The United States’ antitrust policy, if the FTC is to be used as a guide, already has inadvertently strengthened the hand of China. The problems associated with standards technologies are not unlike the problems that the Western democracies are confronting with all emerging technologies. David Delpy, University College London, put it this way:

Western democracies must double down on R&D and strengthen the technological capabilities of business firms. That is a big – but necessary – task. The required R&D must, in the main, be private sector funded. For private sector R&D investment to occur, national policy must make sure that SEP owners receive fair compensation, sufficient to support the business models of those Western firms that generate the technology that ends up as part of the standards technology ensemble. Supporting the SEP licensing process so that it can, in turn, support the R&D necessary for technology development ought to be the critical policy objective of the DOJ, the United States Patent and Trademark Office, and the US National Institute of Standards and Technology. If achieved, it also will support subsidiary goals with respect to competition and consumer welfare.

B. The Changing Geopolitical Landscape

“China Standards 2035,” mentioned in Section III.A, was a galvanizing publication that can serve to remind executives and policy analysts that they need to develop a broader perspective with respect to standards development and standards setting. This initiative builds on “Made in China 2025” and heralds plans and financial and regulatory support for Chinese enterprises, public and private, to take control of the decentralized private (and substantially professionally driven) global standard development process. If successful, China will change the governance of global business, which in turn will augment China’s geostrategic power. There are far-reaching consequences for international business, national security, and competition.

Chinese firms have already obtained substantial representation in the international standard-setting process. Unlike Western representatives, Chinese representatives, whether corporate or government, are held accountable to the nation-state. The Swedish Institute of International Affairs recently noted:

The report further noted that China’s state-directed approach to standards development “radically breaks with both the U.S. and European approaches that are both industry driven.” Other sources draw attention to China, noting:

As a result, one might say that China is “inventing patents,” in the sense that it is diligent about filing for both minor and major inventions. It also is very active at standards-setting meetings and on standards-setting governance.

Unfortunately, there is sometimes limited sophistication in understanding what is going on in the global technology marketplace. Some of this flows from the misreading of patent statistics.

Citing patent analytics company iPlytics, an article in the Wall Street Journal recently noted that companies from China own “36% of all 5G standard essential patents” and that “U.S. firms including Qualcomm and Intel hold just 14%.” The article went on to state:

This statement could be true only if the quality of Chinese patents is equivalent to or better than Western companies’ patents, or if the infringing sales of Western firms are greater than that of Chinese firms. The licensing jurisdictions also need to be similar for such equivalences to be drawn.Footnote ¹⁴ The famous quote “not everything that can be counted counts, and not everything that counts can be counted” seems relevant in this context. However, it would be very imprudent to assume that all Chinese patents are valueless.

Patent statistics can be misleading. Regardless, the United States and its allies – including other liberal democracies, particularly Sweden, Finland, South Korea, and Japan – still maintain a fragile lead, even as the trends do not favor liberal democracies. Policy mistakes now could lead to the rapid dissipation of this fragile leadership by the liberal democracies, with very negative knock-on effects for the US economy and competition. In the next section, I step back and review the context in which SEPs need to be understood – at least with respect to mobile wireless.

C. The 5G Technological Ensemble

The mobile wireless industry has a remarkable track record of developing continuously evolving and improving interoperable systems technology. GSM, Wideband Code Division Multiple Access (WCDMA), and, more recently, Long Term Evolution (LTE) are examples of successful technologies developed privately and separately, but combined by the 3rd Generation Partnership Project (3GPP), using consensus-driven governance, into a platform with massive economies of scale and scope.

Technology development for 5G occurs in a distributed manner with limited overall end-to-end supervision. A very few companies – such as Qualcomm, Nokia, and Ericsson – work hard to help ensure a high degree of end-to-end operability. Without these special efforts, 3GPP could fail. As an organization, 3GPP does not have its own resources to sponsor the development of “gap-filling” technology that, on a standalone basis, may not be financially viable. The real contributions of individual members are hard to calibrate and are not measured merely by counting the number of technical contributions made or patents declared by particular companies.

With 5G, 3GPP has the task of governing a collaborative effort among hundreds of different entities with different interests and incentives. Governance comes in at the time technologies are to be considered for inclusion in a standard. It oversees an iterative, nonlinear, consensus-based approach to technology selection and resulting standards development – systems engineering managed privately and in a decentralized manner. It has worked well, in part because the professionals involved are engineers. Historically, an engineering culture and commercial considerations dominated, and politics were held at bay. Members collectively (by vote) decided on the best technologies to go into a “standard” or new technological ensemble. This may change as Chinese national politics intervene.

Participating firms need confidence that each technology advanced for consideration is robust, has been or will be tested, and can be manufactured, and that the requisite software and applications support will be available. Sponsors of technology then are required to demonstrate that the technology is or can be commercially viable. Hence, by the time that patented technology becomes embedded in the standard, it already has undergone an early assessment as to commercial viability. Licensing executives need to understand this process, as it indicates that patents that are “truly essential” have in all probability passed a litmus test of commercial viability, and thus are likely to have value if indeed they are truly essential and not just “declared essential” by the patent owner.

Feedback from the validation and testing activities is critical and often leads to further development of the technology and/or changes in specification. This process is shown in Figure 1.1. Steps in validation include review, modeling, prototyping, and “plug tests/plug fests,” where designers of equipment or software using the technology proposed for the standard test interoperability of products and designs with those of manufacturers. As standards go through revisions, multiple firms may submit proposals and work together toward final adoption of the standard.

Figure 1.1. Some elements of standards development

The standard-setting for 5G is a continuous process. Updates are issued periodically. Licensing practices have evolved to support the open interoperable mobile wireless ecosystem, with royalties being set in the marketplace via negotiation at levels sufficient to encourage at least a few companies to make the large investment required to develop new 5G technologies.Footnote ¹⁵

D. Technological Ensembles and the Global Open Standardization Miracle

As noted earlier, some observers have the naïve belief that standard-setting is just a matter of choosing compatibility/interoperability standards – like choosing between two- and three-pin electrical sockets and plugs. In reality, the situation is radically different.

Standard development is a distributed global R&D activity that takes place under loose 3GPP/ETSI governance.Footnote ¹⁶

It is perhaps helpful to recognize that 3GPP is akin to a special kind of global technology development consortium that functions as the mobile wireless ecosystem’s R&D arm for developing, assembling, and then standardizing foundational technologies.Footnote ¹⁷ Tens of thousands of engineers work on 5G foundational technologies. They are loosely coordinated by SDOs that work with 3GPP. This amazing technology development program – with billions of R&D dollars spent each year – is mainly privately funded, and the technologies are combined in powerful ensembles by engineers operating under rules established by 3GPP/ETSI.

3GPP periodically releases documents incorporating important new advances in the foundational technologies from the research activities of global mobile wireless technology companies. For example, Release 16 was published in July 2020.Footnote ¹⁸ After a release document is published, it usually takes at least a year before cell phone and other subscribers have access to the fruits of the new technology. Infrastructure companies such as Ericsson, Nokia, Samsung, LG, Huawei, and Cisco have to design the technologies into their equipment; and chip and device manufacturers such as Samsung, Apple, LG, Motorola, ZTE, and Huawei have to embed them in new modems and in new generations of their devices. The network service providers also must install the requisite equipment upgrades for benefits to be realized. They do so, however, long after the standards technology is developed, whereas those that develop standards technology do so without the confidence that their technologies will ever be incorporated in the standards or have commercial payoff of any kind.

3GPP epitomizes cooperative global technology development at its apogee. It is a corollary of the US and European-led post–World War II liberal system of cooperative innovation and distributed economic organization. It is the de facto orchestrator of discoveries and inventions relevant to the mobile wireless ecosystem. With China’s rise and expressed desire to dominate standards setting, this delicate organizational arrangement will be put at risk, particularly if the United States does one more antitrust policy zigzag.Footnote ¹⁹

The type of governance that 3GPP affords is unique as to its global scale and scope. It is rule based and consensus driven. Constituencies include technology developers, systems operators, device makers, and governments around the world.

As described in Section IV, patents are important in the mobile wireless cooperative ecosystem, because patent licensing is how the global sharing and financing of new technology is achieved. Proprietary 2G, 3G, 4G, and now 5G technologies, many of them foundational, have been made available via nonexclusive patent licenses to the whole world, generation after generation. This is a major reason for the success of the mobile wireless ecosystem over the past 30 years. It has afforded countless benefits to billions of users, and fueled competition and economic growth in the United States and abroad. The licensing system promotes both interoperability and entry by device manufacturers by providing a necessary suite of technology input solutions.

A. FRAND and the Open Innovation Ecosystem

The mobile phone industry was in its infancy when ETSI was founded. The focus then and now is on what, in modern terms, we think of as creating a robust innovation ecosystem.

In what follows, I give consideration to these issues, with specific reference to ETSI policy. I also consider the cost of error and elaborate the point that under-rewarding the patent holder of an enabling technologyFootnote ²³ has very high societal costs and should be avoided.

From the outset, ETSI recognized the need for a forward-looking approach to technology development on mobile wireless. The original architects of ETSI’s IP policies sought a “balancing of the interests” of technology contributors (patent owners) and implementers.

ETSI started as a European governmental initiative to assemble a broad set of actors committed to fairness and benefits to the broader telecommunications sector (ecosystem) and consumers. This broad constituency is still apparent today and includes chipset designers and fabricators, handset and base station makers, cellular service providers, app developers, and, of course, consumers.

The standards development system was not designed to favor one constituency over the others, or downstream over upstream. Indeed, initial versions of the ETSI IP policy that did not attract technology developers were rejected in favor of versions that yielded “balance.” When standards technology contributors enable so much of the subsequent downstream innovation, it is critical that technology developers not be shortchanged. This conclusion is not only consistent with ETSI IP policy, but also economically desirable and therefore entirely reasonable from a public policy perspective.

SDOs require that, before technologies are accepted into a standard, members that own patented technologies are “prepared to grant irrevocable licenses on fair, reasonable and non-discriminatory (‘FRAND’) terms and conditions under such IPR.” What is “fair and reasonable” (FR) and what is “nondiscriminatory” (ND) often raise questions. In this chapter, I will address only the FR aspect of FRAND, in the context of ETSI.

There is little doubt – and Dr. Bertram Huber, an ETSI founder representative, confirms this viewFootnote ²⁴ – that ETSI was concerned with establishing a vigorous standards process to support the development of a robust telecommunication industry in Europe and around the world. ETSI requires FRAND commitments from its technology contributors, with the expectation that implementers would take a license under FRAND terms.

B. Patent Holdup: A Theory in Search of a Problem

As noted, a vigorous mobile telecommunication industry requires a robust innovation ecosystem. Various parties and occasionally antitrust regulators have clumsily tried to undo ETSI’s rules with respect to FRAND issues by advancing the concept of “patent holdup.”Footnote ²⁵ Patent infringement by implementers is excused and even encouraged by a “fig leaf” in the form of this antitrust theory of “patent holdup”: “holdup” because the implementer supposedly only knows the royalties they must pay after they have committed capital. The theory chooses to disregard the R&D investments made by the technology developers.

By way of background, consider the sequencing of investment in the development and deployment of standardized technology. Figure 1.2 shows the sequencing of investments, which has important implications for licensing dynamics. Stage one investments are made having no guarantees that they will be successful; and even if it is technologically successful to some degree, the discoveries may not be good enough to go into the 5G standard. The fact that R&D to develop foundational technologies takes place before the equipment and device makers invest puts the licensor in a weak bargaining position with respect to the implementers/licensees. Moreover, there is no guarantee that the technology will be adopted over competing alternatives. This position is amplified by the twin facts that patents are not self-enforcing and injunctions nearly impossible to come by in the United States post-eBay. Enforcement requires a court of law to back up the patent owner with an injunction or something similar. Otherwise “strategic” or “unscrupulous” putative licensees will engage in the similar sounding – but essentially inverse – practice of “holdout.” As Makan Delrahim, then the head of the DOJ Antitrust Division, noted:

Figure 1.2. Sequencing of development and deployment of wireless infrastructure technology

Put differently, implementers can “hold out,” not take a license, and try to dodge paying royalties. To explore these issues further, one must examine in more detail FRAND issues

The mischief comes from implementers, aided and abetted by theoretically oriented academic economists whose models appear to have impacted antitrust enforcement agencies. Academic economic models of “patent holdup” ignore sunk R&D investments by technology developers. They also ignore longer-term business model issues associated with funding R&D. These proclivities create a penchant for seeing royalties as SEP “taxes” and not precious fuel to support R&D.

The intellectual history of patent “holdup” theory is checkered. In my view, in the context of SEPs, it is a theory in search of a problem. It has been an economically damaging red herring that has compromised licensing activities and US technology leadership.

The first (mis)application of the holdup concept to the realm of patents was a paper published by Carl Shapiro in 2001.Footnote ²⁷ Some IP scholars in the legal academy subsequently became aware of, and actively advanced, these patent holdup theories. Meanwhile, other scholars and practitioners actually close to the world of licensing understood this work to be theoretical musing and little else. There was considerable surprise when it was taken seriously by antitrust agencies and some courts.

The assumptions of rampant opportunism and guile by upstream technology providers lie at the heart of holdup theory.Footnote ²⁸ Needless to say, there is no empirical support for either assumption in the context of patents. The theory also assumes that patents are self-enforcing (that is, there is an injunctive right appended to every patent exercisable at the discretion of the (SEP) patent owner), despite eBay. This constellation of assumptions is quite fanciful.

The theory – and theory is all it is – is sometimes articulated in terms of a patent owner promising to one or more implementers one rate, and specific investment is made by an implementer on the basis of that promise. The narrative is that patent owners subsequently, without good reason, strategically increase royalty demands once the implementer is locked in and committed to downstream investments. There is no evidence that patent holdup has ever occurred.

More commonly, the patent owner promises to make licenses available on FRAND terms, but without specifying in great detail until later on which rates it would seek to charge. Moreover, the right to enjoin, when it exists, requires the action of a court. There are, of course, reputational risks that technology developers would have to take should they act egregiously. Needless to say, there are contractual protections for implementers, which are, according to the law of many jurisdictions, third-party beneficiaries of the agreement between the SDO and the patent owner to set royalties that are FRAND.Footnote ²⁹

Technology developers, by contrast, can negotiate only after they have sunk their R&D investments and after their technology has survived a selection process to get into the standard. Thus, technology developers are at least as vulnerable to “reverse holdup” or, as it is more commonly known today, “holdout” by implementers, as implementers are vulnerable to hold up by developers.Footnote ³⁰ Indeed, given that implementers could, but rarely do, seek licenses before they start making standards-compliant products, and the length of time that it takes licensors to conclude agreements with industry players, and the considerations noted below arising from the nature of the enforcement of patent rights, the dangers of holdout are quite real and substantially underestimated.Footnote ³¹ Although the theory of holdup historically has been advocated more vocally, it is holdout that is the greatest risk to licensing of ETSI standards.

Nevertheless, numerous implementers and some nation-states making standards-compliant products have found it convenient to claim to be victims of “holdup.” China, for example, has also tried at the World Trade Organization (WTO) to have all SEPs declared barriers to trade.Footnote ³²

C. Patent Holdout

Fortunately, in recent years, policy concern over holdout has received some attention. As noted earlier, Makan Delrahim, the former head of the DOJ Antitrust Division, made considerable progress in helping develop the understanding that “collective holdout” behaviors in standard-setting are more pernicious and likely than unilateral holdup by SEP holders.Footnote ³³ The recognition that holdup is a red herring is what may have led Contreras to argue that:

From an empirical perspective, Heiden and Petit note the emergence of a “long tail” of implementers or micro-vendors that are individually small but collectively account for a reasonable share of industry revenue, and that are not licensed.Footnote ³⁵ Many of these implementers are based in China. They note in this context that “a systematic patent trespass effect can be deemed to occur when 30% or more of a relevant market is unlicensed.” They relate this to a collective action problem: “why take a license if your competitors do not?” They note that the “systemic effect of patent trespass is primarily experienced through the impact on the technology market through the development of consensus-based standards.”Footnote ³⁶ Heiden, Padilla, and Peters note the presence of a similar “collective action” problem resulting in widespread holdout in the IoT sphere.Footnote ³⁷

These empirical observations echo the findings of Judge Essex of the US International Trade Commission (as summarized by Renaud, Wodarski, and Badin):

In summary, then, the very non-self-enforcing nature of patent rights directly indicates why holdout rather than holdup is a problem that we should expect to see in licensing SEPs. My own experience with examining the smartphone licensing landscape in the context of litigation, and the empirical observations of other commentators, supports this. Royalty revenues are a small share of both the overall value-added from mobile telecommunications and smartphone implementers’ revenues.Footnote ³⁹ This calls into question the predictions of holdup theory and is consistent with the reality that holdout is an important characteristic of the licensing landscape today.

The reason that holdout is a present, and perhaps growing, danger is because it is a profitable strategy for implementers. It is profitable in part because of the weakness of injunctive relief, the fragmentation of the patent enforcement landscape at the global level, and the lack of corrective mechanisms in damages and license awards by courts.

D. Holdup versus Holdout: Time to Close the Debate

Empirical studies have established the irrelevance of holdup theories. Notwithstanding this fact, implementers remain willing and eager to advance such theories. At the same time, it is now widely accepted that contractual mechanisms to redress holdup are available, should it ever occur.Footnote ⁴⁰

Galetovic, Haber, and LevineFootnote ⁴¹ provide a sophisticated empirical analysis with regard to the holdup issue. They find that “products that are SEP-reliant have experienced rapid and sustained price declines over the past 16 years,” and that the “prices of SEP-reliant products have fallen at rates that are not only fast relative to a classic hold-up industry, they are fast relative to the patent-intensive products that are not SEP-reliant.” Using a quasi-natural experiment to study the effect of the eBay decision on relative price declines in SEP-reliant versus non-SEP-reliant industries, they also do not find that prices in SEP-reliant industries were more affected by eBay than in non-SEP-reliant industries. If holdup were more of a problem in SEP-reliant industries, one would have expected to see a greater effect of eBay in these industries than in those not driven by SEPs.Footnote ⁴²

The Galetovic et al. findings are unsurprising. The presence of the FRAND commitment, the lack of availability of injunctive relief (particularly in the United States), the repeat-game nature of standardization,Footnote ⁴³ and the bargaining power of many implementers (for example, their ability to prolong litigation) all militate against holdup. Most fundamentally, holdup is unlikely in a setting where the implementer or prospective licensee can use the technology without paying for it, and injunctions are close to nonexistent, except perhaps in European courts. There is no way that the SEP holder or licensor by itself can prevent infringing use. Accordingly, there is a fundamental difference between “ordinary” goods and services, on the one hand, and IP, on the other – a point that Germany’s Federal Court of Justice recently recognized:

The court’s observation is another way of saying that patents are not self-enforcing. A patent holder cannot do what most suppliers of goods and services can do, which is simply to withhold supply to those customers who do not pay for the good or service.Footnote ⁴⁵ Instead, patent holders must resort to costly and time-consuming litigation to enforce their rights. In such litigation, the risks to licensors and licensees are asymmetric. For example, a court decision that is substantially different from the licensors’ position on the FRAND value of its portfolio potentially can have a significant adverse impact on that licensor’s longer-term licensing strategy. Akemann, Blair, and Teece note:

As noted earlier, the antitrust frameworks that employ holdup are static. They take the funding of (foundational) wireless technology for granted. This further undermines their relevance. It is hard not to agree with Barnett that the academy has led the judiciary and policymakers astray.Footnote ⁴⁷

Accordingly, it is time to shut down the antitrust ruse of patent holdup – not resurrect it, as the DOJ’s draft policy statement on remedies for the infringement of SEPs (issued in December 2021 and withdrawn in June 2022) sought to do.Footnote ⁴⁸ Should evidence of holdup ever emerge, the theory can be resuscitated; but it doesn’t deserve time on the agenda now.

We nevertheless are left with the fallout of previous US policy misadventures in the form of: (a) some foreign competition agencies and bureaus using the holdup argument to support mercantilist policies and favor national champions;Footnote ⁴⁹ and (b) a sense that the debate perhaps has come to a stalemate with legitimate arguments and evidence on both sides. The assessment in (b) is inappropriate. The weight of the evidence favors holdout as the problem. Holdup is merely a theoretical possibility remote from real-world situations. On the other hand, many SEP holders must wait years before they can achieve a license with implementers, or else must resort to litigation – a move that carries asymmetric risks for SEP holders, as discussed later – before they are able to obtain any payment. Moreover, the failure to enter license agreements with particular licensees almost always will have negative consequences for the SEP holder’s broader licensing program.

Additionally, through delaying tactics, licensors may be able to extract significant discounts for past use, benefit from statutes of limitations on past damages, and benefit from potential expiry of patents that they have infringed for many years. If the worst outcome for an infringer is that ultimately it ends up paying a FRAND rate on only some portion of its infringing sales, it will have a great deal of bargaining power to bring to negotiations with SEP holders, meaning that infringement and holdout usually will be profitable strategies.

I consider there to be an urgent need in many jurisdictions to bring “balance” to SEP licensing in the cellular mobile space. While important new cases are moving to recognize the key role of technology developers, implementers still can engage in holdout with only distant sanctions to worry about.

One obvious lever is to strengthen the existing injunctive relief regimes to provide the licensee with stronger incentives to negotiate a license.

However, changes to injunctive relief may be practically difficult to achieve. In this context, FRAND royalty determinations in litigations may play an important restorative role too. Critical to this endeavor of bringing balance is the recognition that infringement and holdout must not be profitable. Recalcitrant licensees must not be put on the same footing as those who were (more) willing to sign up for licenses on FRAND terms. Nor must competition enforcement agencies assist renegades in their quest to avoid paying royalties or to crank down FRAND rates to below reasonable levels.

An important question that should be addressed is whether an “unwilling licensee” – one that is not prepared to accept any license terms other than those that it deems to be FRAND – should lose the benefit of the licensor’s FRAND commitment. This means that not only should injunctive relief be immediate from the moment that the licensee’s unwillingness is established – as the UK courts recognized in a recent decision involving Apple and Optis, and as has been common in mainland Europe – but also potentially that royalty rates in damages awards need not be based on assuming that the FRAND constraint applies.

Inherently, a FRAND royalty rate is reasonably linked to the (likely) value contribution of the technology to the product. However, as discussed earlier (and in more detail later), holdout exerts other costs on the licensor, especially on its overall licensing program. Damage awards based solely on the value contributed to the product may not suffice to restore the licensor’s economic position to where it would have been absent the infringing behavior. Nor would they truly address the harm to the licensing marketplace that holdout behavior inflicts.Footnote ⁵⁰

In practice, many licensees will obtain rates that are well within the FRAND range (and perhaps even below it), because the licensor will be prepared to accept low rates to achieve (relatively) quick settlement and avoid litigation. It would be wrong for courts to base rates for infringers on these negotiated rates without recognizing the context in which these rates were achieved. To do so would put the unwilling licensee on the same footing as the willing licensee. Instead, court-awarded rates (whether applied to licenses or past-use damages) at a minimum should be based on the FRAND (benchmark) rate. However, even this may be too lenient a corrective for the problem of holdout.

Further, the nondiscrimination (ND) prong of FRAND should not be invoked as a reason to base awards either on “best prices” or even averages across licensees – the ND prong cannot be interpreted in such a way that nondiscrimination trumps the fundamental idea of balance. The comparison of royalty rates achieved by different licensees is at most relevant for an ND analysis to the extent that differences in royalty rates distort competition.Footnote ⁵¹ If two licensees operate very different business models and aim at very different market segments, then unwarranted distortions to competition are unlikely. Royalty rates paid to individual licensors are a small sliver of the implementer’s overall cost stack, and so differences in these rates paid are unlikely to distort competition.Footnote ⁵²

As noted earlier in Section IV (especially Figure 1.2), the commonplace situation that exists is one in which an implementer begins manufacturing devices and implementing SEPs long before taking licenses to any of them. Many licensors issue notice letters to implementers years after use of the SEPs actually began. It is often at least a couple of years before licenses are agreed. Reasonably often, there is no agreement, and unlicensed use continues unabated. In many cases, perhaps something close to a decade will go by before litigation is launched, and then perhaps one or two further years will pass before any decision – and before any court-awarded damages or court-determined license are made available. In all, a decade or more might pass before the infringer pays anything for its use, by which point a new generation of the standard has replaced the old standard upon which the litigated decision was based!

Even then, the licensee may pay no more than a relatively low rate that is deemed to be FRAND – often on the basis of rates derived from comparable licenses without necessary adjustments to account for the difference between willing licensees and those who force the licensor into litigation. Alternatively, this rate may be based on inherently licensee-friendly formulations such as the “top-down” method of allocating some aggregate reasonable royalty among the different licensors. Either way, a licensee that delays or strings out discussions has little incentive to take a license – at worst, it will have to pay the same FRAND rate that it would have paid anyway. At best, it secures an advantage over its licensed rivals by remaining under the licensor’s radar.

Growing diversity in the nature and geographic loci of implementers and the emergence of new use cases such as those associated with IoT exacerbate the problem. The share of implementers with licenses likely is falling, and at least in some environments – such as the licensing of IoT implementers – a collective action problem is emerging in which widespread infringement may be self-perpetuating.Footnote ⁵³ The source of the collective action problem is simple: Licensed users will see themselves as being competitively disadvantaged relative to unlicensed users.

That such pervasive problems exist is not surprising. The perverse focus on holdup and the introduction of antitrust lawsuits has made these problems worse. Attention to the specifics of licensing negotiations and actual business behavior always has suggested that holdout rather than holdup is the bigger problem.

A. Why Corralling SEP Infringers Is Difficult

The primary issue that should animate competition policymakers and the courts is the threat posed to open consensus-based standards development by holdout behavior. At its root, a holdout licensee will do no worse than a willing one; and it might well do better. That is, it could end up paying nothing or, by negotiating a license late in the day (when most of its sales and profits from using the technology are safely in the past), pay heavily discounted rates.

Similarly, consider a situation in which the licensor offered a license on FRAND terms, and the licensee declined such a license and made counteroffers that were not FRAND or otherwise indicated its unwillingness to take a license on FRAND terms. In this case, even if the licensee were enjoined, it can still avail itself of a FRAND license at some point in the future.

The perennial availability of a FRAND license weakens even the power of any injunctive relief remedy – where such is available. It encourages the licensee to try its luck in the courts rather than negotiate for a license. If the licensee escapes an injunction, it can continue to enjoy the benefit of infringement perpetually or until it forces the licensor into conceding terms that it likes.Footnote ⁵⁴ Alternatively, if the licensee is informed, it may still have the option to have the injunction lifted by accepting a FRAND license – possibly the very same FRAND license it had been offered and turned down before. Nonetheless, the threat of injunction can still have a powerful effect on some recalcitrant licensees.

The present standardization and licensing systems lack sufficient corrective or countervailing forces to prevent this problem from not just entrenching itself but getting worse – as the unlicensed share of the industry increases, the stronger will be the incentives for other firms to resist taking licenses too. Short of some form of enhanced damages, there do not seem to be innate solutions to this problem of potential mass infringement – at least not for implementers who are not major contributors to standards and therefore do not face consequences for past opportunistic conduct.

Worse still, there is some indication that holdout is now something of a norm – for example, very few implementers sign a license anywhere close to their date of first infringement, and most limit the period of past use for which they pay royalties and/or are able to negotiate steep discounts for past infringement. Thus, as delay and discounts related to that delay become a norm across all implementers, the harder it becomes to sanction such behavior through the standard-setting process.

Instead, it seems that the best restraint on holdout is likely to arise from a willingness on the part of the courts (and the competition agencies) to engage with the intent of the FRAND commitment (that is, to recognize fully the centrality of balanced incentives to ETSI’s standardization activities). It also is important that antitrust enforcement agencies keep their distance and don’t provide a helping hand to infringing holdouts.

B. Strengthening Injunctive Relief

The most obvious factor encouraging holdout is the difficulty of obtaining injunctive relief. No US court has granted an injunction in a SEP-related case, at least not since eBay. The situation in Europe is somewhat better than that in the United States, and there is a well-developed framework (Huawei-ZTE) for assessing when injunctive relief is an appropriate remedy.Footnote ⁵⁵ However, even that framework arguably provides too much leeway to the putative licensee or infringer. This can be understood usefully by contrasting the Huawei-ZTE framework to Germany’s “Orange Book” standard.Footnote ⁵⁶

Under Germany’s Orange Book standard, an infringer could avail itself of a competition law defense against an injunction only if it had made an unconditional offer to license either on terms that the SEP holder could not reasonably refuse or at a rate proposed by the SEP holder but subject to review and modification by courts. Under Huawei-ZTE, the licensee is not required to make such commitments. Importantly, the “unconditional” commitment to license under Orange Book means that the acceptance of a license cannot be delayed through challenges to selected patents on validity and infringement grounds. The Orange Book standard was seen as very generous to patentees and was not designed in the context of ETSI SEPs, where the SEP holder has made an explicit FRAND commitment. Nonetheless, some of its provisions may have merit in the context of cellular SEPs. As a practical matter, the licensing of ETSI-related SEPs is almost always at the portfolio level. Portfolios that have been licensed and/or litigated frequently almost certainly contain at least some valid and infringed patents. In this context, the leeway of licensees to resist taking portfolio licenses by challenging specific patents on validity and infringement grounds (as allowed under Huawei-ZTE) is an avenue for delay and another mechanism by which licensees are able to raise SEP holders’ enforcement costs. The practical upshot is that under the Huawei-ZTE regime, holdout has become easier, and SEP-related injunctions have become harder to obtain.

The UK Supreme Court in Unwired Planet International Ltd v. Huawei Technologies (UK) Co. Ltd devised the concept of a “FRAND injunction.”Footnote ⁵⁷ Injunctions were available only at the point that the licensee turned down a FRAND determination from the court. In practice, this allows licensees to go all the way up to the “FRAND trial” without any penalty for infringement – in the United Kingdom, the trial to determine FRAND terms would occur after a typically lengthy and costly process of validity and infringement trials, as well as trials on separate, discrete issues. Again, given this potential for delay and given that the licensee’s worst-case scenario is a court-determined FRAND license, the threat of an injunction in the UK court may not be a powerful motivation for implementers to negotiate a license with urgency. Indeed, some implementers may be quite happy to accept an injunction in the relatively small UK market if the alternative is to avoid a global license agreement on terms determined by a UK court.Footnote ⁵⁸

C. Adjusting Damages to Disincentivize Holdout

Courts should recognize that holdout creates significant economic harms for the SEP holder in question, to the innovation ecosystem, and for the licensing marketplace. The problem in individual litigations is typically that a given SEP holder has failed to secure a license after a prolonged period of infringement by the implementer. As noted, the harm in that particular case is not just the cost of delayed payment to the patent owner, but the indirect harm to the SEP holder’s licensing program.

There are clear externalities in licensing – securing a given license can confer credibility and momentum for the SEP holder’s broader licensing efforts. Conversely, failure to secure a license can damage progress with other would-be licensees. In particular, some implementers – especially those in product segments such as IoT or among implementers focused on some emerging markets and on China – will seek to tie either the terms of a license (for example, payment) or the very signing of a license to the SEP holder’s success in signing on other implementers that they perceive as being in the same segment. There may be a broader impact on the licensing marketplace as a whole. Akemann, Blair, and Teece explain how widespread infringement begets more infringement, thus creating a “bandwagon effect”:

How courts handle issues such as damages or the terms of FRAND licenses can thus have an impact in terms of not just alleviating economic harm arising from prolonged infringement, but also correcting the distortions in the wider licensing marketplace that arise from allowing such conduct to persist. The actions that courts take today will dictate not just future litigation outcomes, but future negotiations in the marketplace – which always happen in the “shadow of the court.”

D. Limiting the Availability of FRAND

Consider an SEP holder that has been attempting to negotiate a license with an implementer for several years and finally has brought the matter to litigation. If the evidence suggests that the putative licensee essentially had no interest in negotiating a license on FRAND terms, is it adequate that the redress available is a license on FRAND terms, and damages for past infringement on FRAND terms? This is problematic for two reasons: (a) It potentially puts the litigious infringer on the same footing as more willing licensees; and (b) it actually does not restore the SEP holder into the position it would have been had the infringement never taken place, as the SEP holder will not be compensated adequately for the harm it has suffered as a result of the infringement. While in principle the harm caused by the delay in taking the license might be quantifiable, the harm to the wider licensing program may not be so readily quantifiable. With respect to addressing the harm caused just by the delay to taking the license, courts should be prepared to address this harm and to do so in an economically robust way – for example, instead of using statutory interest rates in damages awards, courts could use the licensor’s cost of capital or some other measure of economic opportunity cost to address the issue.

More importantly, instead of giving the licensee the choice of eventually taking a FRAND license, once a licensee has been found unwilling, the most obvious corrective is to withdraw the option of a FRAND license. The threat of being found unwilling and thus losing an entitlement to a FRAND license is likely more potent than the threat of being enjoined (which will only apply in a single jurisdiction anyway) and then being able to claim a FRAND license to lift the injunction. (If courts across different jurisdictions consistently applied this logic, then it would also prevent the situation wherein the implementer can swallow an injunction in a less important jurisdiction and then prevail upon another court to award it a FRAND license.)Footnote ⁶⁰

These proposals might seem radical to some and might push courts into territory that seems controversial. After all, there is nothing explicit in the ETSI IP policy that suggests that the FRAND commitment applies regardless of the licensee’s willingness; but nor is there an explicit provision that limits its application in the case of an unwilling licensee. To the extent that above-FRAND awards might contain a deterrent element, they may be seen by some as legally difficult to justify, but they are economically very easy to justify.Footnote ⁶¹ Nonetheless, at least the broader concept of restoring the licensor’s economic position by recognizing the harm it suffers from infringement fits in with the idea of restorative or equitable damages, rather than being punitive.Footnote ⁶²

A minimalist measure that would at least somewhat restore the balance is that courts should not give unwilling licensees the same terms as willing ones. Quite often, the determination of FRAND rates is based (at least partly) on a review of the licensor’s “comparable” licenses. In many instances, the rates that the licensor negotiated with the licensee might reflect concessions that were made in the context of meaningful negotiations. The licensor has an interest not just in securing the best “rates” but in ensuring that its licensing program maintains momentum (given that concluding license agreements lends impetus to concluding other licensing agreements – that is, there are externalities at work). Clearly, such a calculus does not apply in the case of the unwilling licensee, and equally clearly, there is a compelling policy rationale for not putting unwilling and willing licensees on the same footing. Put another way, while a range of rates may be consistent with FRAND, the unwilling licensee ought not be entitled to the “best” rate in that range, or even the average rate in that range. In fact, it would still be quite accommodating to give it a rate that is at the top end of the FRAND range.

E. The Role of Competition Policy

With their endorsement of holdup theories, US competition agencies sometimes have eschewed an evidence-based approach to antitrust and wittingly or unwittingly aided and abetted patent holdout, thereby compromising the development of foundational technology that supports innovation and drives dynamic competition. The primary beneficiaries of this strategy are new entrants in China that end up receiving a discount, and/or escape royalty payments for many years – and often permanently.Footnote ⁶³

Competition policy needs to favor the future and embrace innovation and business models that support innovation in the key enabling technologies. They must support innovation both upstream and downstream. Avoiding the temptation to resurrect patent holdup arguments is the most concrete step that can be taken to help innovation and competition. Challenging the open innovation model simply plays into the hands of those that might prefer vertical integration of upstream wireless technology and downstream devices.